不同品种牛肉新鲜度光谱检测模型的维护方法

目的 研究不同品种牛肉新鲜度挥发性盐基氮 (TVB-N)定量光谱检测模型的修正方法, 对模型进行维护, 从而提高光谱校正模型的通用性。方法 采用基于SPXY算法的代表性样本添加及光谱吸光度值修正两种方法对黄陂黄牛的TVB-N含量预测模型进行修正。结果 两种方法修正之后的模型使恩施水牛样本的TVB-N预测相关系数R从0.021分别提高至0.919、0.851, 预测误差从8.347分别降至2.182、2.872。结论 修正之后的黄牛模型可实现对黄陂黄牛和恩施水牛两个品种新鲜度TVB-N含量的定量检测, 样本添加的方法优于吸光度值修正方法。

Maintenance methods of freshness detection model for beef of different species based on spectrum

Objective To study the maintenance methods of freshness (total volatile basic nitrogen, TVB-N) detection model for beef of different species based on spectrum and improve the generality of the calibration model. Methods Model updating by adding new typical samples based on SPXY algorithm and absorbance modification were adopted to improve the model applicability of Huangpi cattle. Results For model updating, the model correlation coefficient of Enshi buffalo increased from 0.021 to 0.919, with the RMSEP from 8.347 to 2.182. And for absorbance modification, the two parameters were improved to 0.851 and 2.872, respectively. Conclusion Huangpi cattle model after maintenance had a good applicability for both Huangpi cattle and Enshi buffalo. As for the two maintenance methods, model updating by adding new typical samples had a better performance than the absorbance modification.